Add minimal b-jet trigger matching HLT content to DAOD_PHYS

Supersedes !73825 (closed)

Stripped down to just the content additions to ease getting this in for end of Sep. Larger overhaul of trigger content to come later.

This makes adds HLT b-jet trigger matching content (no L1 jets here) for Run 2 and Run 3 PHYS. The content and sizes evaluated on 1k events from a random AOD input file per year are as follows:

• Run 2
  • HLT jet branches: pt,eta,phi,m for
    • 2015/16: EFJets, SplitJets, a4tcemsubjesFS (untagged jets) -- 400B - 1kB per event, mostly the plain jets -- already reduced to 1/4 of their original size
    • 2017/18: GSCJets, SplitJets, a4tcemsubjesISFS (untagged jets) -- 1-2kB per event
    • We should be able to remove the a4tcemsubjes(IS)FS jets at a later stage.
  • HLT b-tagging branches
    • HLTBjetFex, only MV2c[00,10,20]_discriminant -- 20-40B per event
• Run 3
  • HLT b-jet branches: pt,eta,phi,m for HLT_AntiKt4EMPFlowJets_subresjesgscIS_ftf_bJets, 95B per event
    • Possibly useful to check match against jets from navigation (@lshi is this true, or should we just leave out?)
  • HLT b-tagging branches: pu,pc,pb for the recommended tagger in each year, 40B per event

So for Run 3 this is very compact, a little more than 100B/event. For Run 2 we would expect to eliminate the a4tcem jets at a later date, cutting this down to approx 600B/event but in the first round of the navigation conversion (assuming it goes in) we may not have fully validated jet matching from the slimmed navigation four-vectors.